Configure Hermes Models and Providers

---
name: configure-hermes-models-and-providers
description: Choose and configure Hermes Agent model providers with a fallback and verification plan.
---

# Configure Hermes Models and Providers

Use this skill when setting up or troubleshooting Hermes Agent model access.

## Inputs
- `preferred_provider`: provider the user wants to use, if any
- `budget_priority`: lowest cost, best quality, fastest, local-first, or balanced
- `model_tasks`: coding, research, long-context, vision, automation, or chat
- `existing_credentials`: which providers are already configured, without revealing secrets

## Procedure
1. Recommend one primary provider and one fallback provider.
2. Explain where credentials should live locally, not in chat.
3. Suggest model choices by task: coding, general chat, long context, auxiliary vision/compression, and cheap routing.
4. Include a smoke test: run a simple single-query chat, then a tool-using request.
5. If a provider fails, inspect config, environment, auth status, and provider-specific error messages before switching models.

## Output format
### Provider recommendation
### Model map by task
### Local configuration checklist
### Smoke tests
### Troubleshooting notes

## Rules
- Do not invent provider availability or pricing; check current provider docs or local config when needed.
- Do not expose keys in logs or user-visible output.
- Prefer provider-agnostic wording so users can adapt this to OpenRouter, Anthropic, OpenAI, local models, or custom endpoints.